Data Science Intern

Proofpoint•Sunnyvale, CA

1d•Onsite

About The Position

Join our AI team building next-gen Digital Communications Governance features, including LLM tuning, vectorization/embeddings for semantic search, automated redaction, and Supervision reviewer recommendation workflows for investigations and compliance. What you’ll do Prototype and evaluate embedding + retrieval pipelines (chunking, indexing, reranking) for investigation-grade semantic search Assist with LLM tuning approaches (prompting, lightweight fine-tuning, eval harnesses) for classification, summarization, and recommendation tasks Build/extend redaction models/rules (PII/entity detection), and measure precision/recall tradeoffs Design experiments, create metrics, and document findings; ship working code and repeatable notebooks/pipelines

Requirements

Strong Python and practical ML/data wrangling (pandas, numpy; PyTorch or similar)
Solid understanding of NLP/LLMs: embeddings, retrieval-augmented generation (RAG), evaluation methods
Experience with at least one: vector databases (FAISS, Milvus, Pinecone, Elasticsearch/OpenSearch vector), or building ANN search
Comfort with experimentation: offline evaluation, A/B-style thinking, error analysis
Clear communication and ability to deliver in short iterations

Responsibilities

Prototype and evaluate embedding + retrieval pipelines (chunking, indexing, reranking) for investigation-grade semantic search
Assist with LLM tuning approaches (prompting, lightweight fine-tuning, eval harnesses) for classification, summarization, and recommendation tasks
Build/extend redaction models/rules (PII/entity detection), and measure precision/recall tradeoffs
Design experiments, create metrics, and document findings; ship working code and repeatable notebooks/pipelines